Biological Conservation 191 (2015) 428–435 Contents lists available at ScienceDirect Biological Conservation journal homepage: www.elsevier.com/locate/bioc Reintroduction of Tasmanian devils to mainland Australia can restore top-down control in ecosystems where dingoes have been extirpated Daniel O. Hunter a,⁎,ThomasBritzb, Menna Jones c,MikeLetnica a Centre for Ecosystem Science, University of New South Wales, NSW 2052, Sydney, Australia b School of Mathematics and Statistics, University of New South Wales, NSW 2052, Sydney, Australia c Department of Zoology, University of Tasmania, Tasmania 7001, Hobart, Australia article info abstract Article history: Restoring missing ecological interactions by reintroducing locally extinct species or ecological surrogates for ex- Received 16 February 2015 tinct species has been mooted as an approach to restore ecosystems. Australia's apex predator, the dingo, is sub- Received in revised form 21 July 2015 ject to culling in order to prevent attacks on livestock. Dingo culling has been linked to ecological cascades Accepted 24 July 2015 evidenced by irruptions of herbivores and introduced mesopredators and declines of small and medium sized Available online xxxx mammals. Maintenance of dingo populations is untenable for land-managers in many parts of Australia owing to their depredations on livestock. However, it may be possible to fill the apex predator niche with the Tasmanian Keywords: Rewilding devil which has less impact on livestock. Devils once occurred throughout Australia, but became extinct from the Trophic cascades mainland about 3000 years ago, but are now threatened by a disease epidemic in Tasmania. To explore the fea- Predator diversity sibility of reintroducing devils to mainland Australia we used species distribution models (SDMs) to determine Tasmanian devils if suitable climatic conditions for devils exist and fuzzy cognitive mapping (FCM) to predict the effects of devil Dingoes reintroduction. Based on devils' current distribution, our SDM indicates that suitable areas for devils exist in south-eastern Australia. Our FCM examined ecosystem responses to predator-management scenarios by ma- nipulating the abundances of devils, dingoes and foxes. Our FCMs showed devils would have cascading effects similar to, but weaker than those of dingoes. Devil introduction was linked to lower abundances of introduced mesopredators and herbivores. Abundances of small and medium sized mammals and understorey vegetation complexity increased with devil introduction. However, threatened species vulnerable to fox predation benefited little from devil introduction. Our study suggests that reintroducing ecological surrogates for apex predators may yield benefits for biodiversity conservation. © 2015 Elsevier Ltd. All rights reserved. 1. Introduction (Donlan et al., 2006). Reintroducing strongly interactive species may also be desirable to curb the effects of invasive species, or to fill vacant All species interact with other species both directly and indirectly, niches in novel ecosystems that are dominated by suites of invasive through mechanisms such as predation, competition, mutualism, facili- species. tation and herbivory. The outcomes of such interactions are among the Large carnivores are often keystone species because they typically major factors that shape ecological communities. Species are described occur at low densities and may exert top-down control on ecosystems as strongly interactive when their absence leads to significant changes through their predatory and competitive interactions with herbivores in ecosystems and as keystone species when the strength of their inter- and smaller predators (Ripple et al., 2014). Because top predators kill actions is disproportionate to their abundance (Soule et al., 2003). their prey and frequently kill their competitors as well, the most obvious Increasingly, ecologists are realizing that the effects of strongly inter- population-level effects of top-predator removal are increases in the active species could be used to manipulate ecological processes and abundance of large herbivores and smaller predators (mesopredators) achieve biodiversity conservation goals (Seddon et al., 2014; Ritchie (Ripple et al., 2014). et al., 2012). Understanding the important role that strongly interactive While the direct predatory effects of top predators are relatively species can play in shaping ecosystems has prompted debate about the easily observed, they typically have strong indirect effects as well. For merits of restoring missing ecological interactions by reintroducing lo- example, numerous studies have found that, in the absence of a top- cally extinct species or ecological surrogates for now extinct species predator, herbivores increase in abundance, reducing the biomass of palatable plants (Ripple et al., 2014). A similar cascade of effects is predicted by the mesopredator release hypothesis (MRH). The MRH ⁎ Corresponding author at: Daniel Hunter, School of BEES, University of NSW, Kensington, NSW 2052, Australia. postulates that reduced abundance of top-order predators results in E-mail address: [email protected] (D.O. Hunter). an increase in the abundance and predatory impact of smaller predators http://dx.doi.org/10.1016/j.biocon.2015.07.030 0006-3207/© 2015 Elsevier Ltd. All rights reserved. D.O. Hunter et al. / Biological Conservation 191 (2015) 428–435 429 (Soulé et al., 1988). Consequently, small prey species that are the one of the reasons why small and medium sized native mammals are so preferred prey of mesopredators may decline in abundance (Prugh abundant there today (Johnson, 2006). et al., 2009). Owing to the amplified interactions of herbivores and Importantly, devils are not as great a threat to livestock as are mesopredators, terrestrial ecosystems from which apex predators dingoes, although their predatory impact is at least similar to that of have been removed tend to be characterised by reduced diversity of foxes with young sheep at risk (Jones et al., 2003), and thus their reintro- small vertebrates and shifts in vegetation composition (Letnic et al., duction is likely to obtain more support from livestock producers than 2012; Ripple et al., 2014), but these trends are not evidenced in all would the maintenance of dingo populations. Since the turn of the cases (Letnic et al., 2012). 21st century, devils have undergone massive population decline due to Australia's largest mammalian carnivore is the dingo (Canis dingo). an epidemic of devil facial tumour disease (DFTD) (Hollings et al., Dingoes arrived on mainland Australia 5000–3500 years before present 2014). Thus, a mainland population of DFTD-free devils would also func- (ybp) (Letnic et al., 2014). Around the time that the dingo arrived tion as an insurance population in the event that the Tasmanian popula- on mainland Australia, two marsupial predators, the Tasmanian tiger tion became extinct as a result of DFTD or another threatening process. It (Thylacinus cynocephalus) and Tasmanian devil (hereon in devil) is conceivable also, that a reintroduction of devils to the mainland may (Sarcophilus harrisii) became extinct from mainland Australia but succeed because two of the purported drivers of devil decline from the persisted on the island of Tasmania which dingoes never colonised mainland during the Holocene, dingoes and hunting by Aboriginal peo- (Letnic et al., 2014). The coincidence in the timing of the arrival of the ple, are now diminished across much of the mainland and thus may not dingo and extinction of the Tasmanian tiger and devil, coupled with the pose a threat to reintroduced devil populations. fact that larger dingoes were likely to be superior competitors in one We use two approaches to evaluate the potential to reintroduce on one agonistic interactions lends support to the idea that dingoes likely devils to south-eastern Australia. Because devils are now restricted to contributed to the extinction of these marsupial predators from main- the island of Tasmania which has a temperate climate, it is possible land Australia (Letnic et al., 2012). This hypothesis is supported by obser- that extant devils may be poorly adapted to the warmer environments vations that devils are easily killed by domestic dogs (Jones et al., 2003). of mainland Australia. Given this possibility, it makes sense to initially It has also been proposed that increasing human population densi- reintroduce devils to areas that are climatically similar to Tasmania. ties and or climate, not dingoes were the main driver of the extinction Thus, we used species distribution models (SDMs) to identify areas of the Tasmanian tiger and devil from mainland Australia (Johnson that are climatically suitable for devils based on their existing distribu- and Wroe, 2003). Recent analysis of mitochondrial DNA reveals tion in Tasmania. that drought associated with a severe El Niño Southern Oscillation cli- Our second aim was to use fuzzy cognitive maps (FCMs) to predict matic event from 3000–5000 ybp coincided with the decline of devil the ecological outcomes of predator management scenarios. FCMs populations (Brüniche-Olsen et al., 2014). Extreme drought may have predict the outcomes of interactions among multiple input species pushed devil populations to sufficiently low levels that a “predator (Ramsey and Norbury, 2009; Dexter et al., 2012) and are a risk-free, pit” (Sinclair et al., 1998) ensued whereby predation by humans and low input method of testing novel management scenarios. The scenarios